CES 2026 opens January 6 in Las Vegas. In past years, the headlines skewed toward novel devices and showcase demos. This year, the agenda signals something different: the week’s highest-value announcements will be about the systems that make advanced technology dependable at scale—compute platforms, edge deployment, industrial automation, connectivity, and security.

That shift matters across the board: senior engineers looking for realistic road maps, directors and vice presidents budgeting for 2026 rollouts, and founders deciding where to place their bets. AI is no longer the “feature.” It is the workload—and the show’s best signals will come from companies explaining how they intend to run it reliably, securely, and profitably in the real world.

Key Dates That Set the Tone

CES is officially January 6–9, but the week effectively begins with Media Days on Sunday and Monday (January 4–5). Those two days concentrate a large share of the week’s top-level narrative: early product announcements, partnership signals, and the first “here is what we think matters” positioning from major brands.

If you only have time for the highest-signal moments, focus on:

• The major semiconductor and computing statements on Monday, before the show floor opens.
• The keynote slate on Monday night and Tuesday/Wednesday morning, which leans industrial and enterprise.
• CES Foundry on January 7–8, which is explicitly designed around AI and quantum themes.

The Big Themes and Announcements to Watch

1) Semiconductors: the “AI everywhere” compute refresh

The most concrete, pre-announced hardware signal coming into CES is a major client and edge processor launch. Intel is using CES to stage the global launch of its Intel Core Ultra Series 3 processors (Panther Lake), positioning them not only for PCs but also for “edge solutions” and new AI experiences. In other words: this is not just an enthusiast PC story; it is part of the broader “inference moves outward” story.

AMD’s CES keynote is similarly framed around AI solutions “from cloud to enterprise, edge and devices.” Pay attention to whether the company draws a clear line from data center road maps to edge inference deployments, and whether it provides credible software and fleet-management narratives—not just silicon.

NVIDIA, for its part, is coming into CES with the language of full-spectrum AI: not only accelerated computing, but also robotics, industrial simulation, and inference tooling. That is a direct infrastructure posture: what matters is less the single chip and more the end-to-end deployment model.

What to listen for, across all the silicon announcements:

• Performance per watt and thermal realities, especially for edge and embedded use.
• The “software contract” vendors are implicitly offering—how models are deployed, updated, monitored, and secured in the field.
• Interconnect and memory narratives: the compute bottlenecks are increasingly architectural, not simply clock-speed stories.

2) Manufacturing and industrial automation: digital twins move from talking point to operating model

A defining infrastructure signal on the keynote schedule: Siemens is using its CES keynote to describe how AI, digital twins, and automation are reshaping manufacturing and infrastructure. That focus is consistent with what the director-level audience is demanding: measurable outcomes—throughput, downtime reduction, quality, safety—not abstract futurism.

This industrial push is also reinforced by NVIDIA’s CES programming: its materials explicitly frame “industrial AI” as a revolution, and place Jensen Huang alongside Siemens leadership in that narrative. Taken together, this is a strong indicator that CES 2026 will treat industrial AI as core—not peripheral.

What to watch here:

• Whether “digital twin” is presented as an engineering workflow (design-to-operations continuity), not a marketing metaphor.
• The extent to which these systems are offered as interoperable components rather than closed environments.
• How industrial AI claims are validated: reference architectures, benchmarks, safety cases, and real deployment examples.

3) Robotics: from spectacle to commercialization

Robotics is no longer confined to curated demos; it is increasingly tied to labor realities, safety mandates, and competitive production economics.

Hyundai Motor Group is bringing an explicit AI robotics strategy to CES under the theme “Partnering Human Progress,” with an emphasis on human-robot collaboration and manufacturing innovation. It is also positioning Boston Dynamics’ new Atlas for a public debut at CES, and describing “software-defined factory” ideas as part of the commercialization plan.

On the consumer-facing side, LG is using CES to introduce a home robot—LG CLOiD—and to frame it as a home assistant designed for real household tasks. Whether you view home robotics as imminent or aspirational, the larger signal is important: robotics is being presented as a product category that requires a credible platform story (sensing, dexterity, safety, updates, and service), not just a prototype.

What to listen for:

• Safety and reliability framing (especially around human-robot collaboration).
• The training and update model: how robots improve over time without turning deployments into continuous science projects.
• The degree of openness: APIs, integration with industrial systems, and support models.

4) Connectivity and telecommunications: Wi-Fi 7, 5G, and the edge reality check

The CES conference program includes a dedicated “Mobile Technology and Connectivity” track, and the schedule explicitly includes sessions built around “Wi-Fi 7, 5G and beyond.” For IoT and edge AI, that matters because connectivity is now inseparable from the AI deployment story: model updates, telemetry, remote management, and security posture all depend on dependable networks.

NVIDIA’s CES materials even highlight telecommunications explicitly within its AI software portfolio, a reminder that AI workloads are increasingly seen as relevant to radio access networks, edge compute, and network operations.

What to watch:

• Practical network stories: deterministic performance, private networks, and measurable improvements in coverage or latency.
• Where inference happens (on-device, at the edge, or centralized), and what that implies for bandwidth and costs.
• Evidence of real carrier, vendor, or enterprise deployments—not just lab demonstrations.

5) Cybersecurity: the edge cannot scale without trust

CES 2026’s official conference program explicitly calls out cybersecurity as a core topic, and that is not window dressing. As AI spreads into endpoints, industrial systems, and vehicles, the attack surface expands—and the cost of compromised systems increases.

Expect cybersecurity stories to show up in three places:

• Device identity, secure provisioning, and lifecycle management for IoT fleets.
• Safety and security requirements for robotics and mobility.
• Secure pipelines for model delivery and updates.

The best CES security announcements tend to be specific: concrete architectures, standards alignment, and verification approaches rather than generic reassurances.

6) Quantum computing: more pilots, fewer slogans

Quantum will not “arrive” at CES as a consumer product. What will matter is whether the industry is moving from generalized promise to specific, testable pilot programs.

Two CES mechanisms make this particularly visible in 2026:

• CES Foundry, positioned as a home for AI and quantum innovation, with programming that explicitly spans enterprise AI trust and quantum applications.
• A dedicated “Quantum Means Business” conference track aligned with the CES program.

The signal to watch: credible problem statements and timelines. Who is piloting what, in which domains, and what counts as success?

What to Expect From Jensen Huang’s CES Presentation

Jensen Huang’s CES event is framed in unusually direct terms: “what’s next in AI,” and how NVIDIA’s latest solutions drive innovation and productivity across industries. Based on NVIDIA’s own CES materials—and the way its schedule is constructed—expect Huang to emphasize several themes that align with infrastructure, not novelty.

1) AI as a full deployment model, not a single product.

NVIDIA is marketing CES around an end-to-end proposition: hardware and software that support real deployments. Expect Huang to connect model development, inference, and operational management into a single narrative.

2) Physical AI and robotics.

NVIDIA is explicitly programming CES Foundry content around “physical AI” and robotics, and its CES materials emphasize robotics and edge AI topics. That makes it likely that Huang will reinforce the shift from “AI that talks” to “AI that acts,” especially in robotics, autonomy, and industrial settings.

3) Industrial AI and digital twins.

NVIDIA’s CES schedule places Huang in an industrial AI discussion tied to Siemens leadership. That is a strong indicator that manufacturing, simulation, and automation will be treated as first-class AI domains—where success is measured in uptime and throughput, not only model quality.

4) Telecommunications and edge inference as strategic terrain.

NVIDIA’s CES materials explicitly reference telecommunications within its software portfolio. If Huang addresses telecom directly, it will likely be in the context of AI running at the edge of networks and inside critical infrastructure.

How Tech Scope Connect Will Cover CES

Tech Scope Connect launches Thursday, January 8—right as the show transitions from initial headline announcements to the deeper technical and commercial details that determine what will actually ship and scale. Our CES coverage will concentrate on the decisions that matter for 2026 programs:

• Which announcements materially change compute, deployment, or lifecycle economics.
• What is ready for production, and what is still positioned as a concept.
• How vendors are treating security, fleet operations, and interoperability.

CES remains a proving ground. This year, the proof will be infrastructure.